arrow
返回

DIODE: Dilatable Incremental Object Detection

delete2023-04-01
delete3
delete
OA
AI
C
Can Peng *
K
Kun Zhao
S
Sam Maksoud
T
Tianren Wang
B
Brian C. Lovell
DOI:10.1016/j.patcog.2022.109244delete
delete原文链接
delete分享
delete收藏
查看原文
摘要

摘要

En 中文
To accommodate rapid changes in the real world, the cognition system of humans is capable of continually learning concepts. On the contrary, conventional deep learning models lack this capability of preserving previously learned knowledge. When a neural network is fine-tuned to learn new tasks, its performance on previously trained tasks will significantly deteriorate. Many recent works on incremental object detection tackle this problem by introducing advanced regularization. Although these methods have shown promising results, the benefits are often short-lived after the first incremental step. Under multi-step incremental learning, the trade-off between old knowledge preserving and new task learning becomes progressively more severe. Thus, the performance of regularization-based incremental object detectors gradually decays for subsequent learning steps. In this paper, we aim to alleviate this performance decay on multi-step incremental detection tasks by proposing a dilatable incremental object detector (DIODE). For the task-shared parameters, our method adaptively penalizes the changes of important weights for previous tasks. At the same time, the structure of the model is dilated or expanded by a limited number of task-specific parameters to promote new task learning. Extensive experiments on PASCAL VOC and COCO datasets demonstrate substantial improvements over the state-of-the-art methods. Notably, compared with the state-of-the-art methods, our method achieves up to 6.0% performance improvement by increasing the number of parameters by just 1.2% for each newly learned task. (c) 2022 Elsevier Ltd. All rights reserved.
Keyword:
Incremental learning
Object detection
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

Pattern Recognition 封面图
Pattern Recognition
IF:
7.6
论文数:
1.3W
被引数:
4.5W

机构

U
University of Queensland
学者数:
5.0W
论文数: 5.1W
被引数: 9.2W
引用论文

引用论文

err分享
err收藏
The Pascal Visual Object Classes (VOC) ChallengePascal视觉对象课程 (VOC) 挑战
err2009-09-09
err9.0K
PREAI
errEveringham, Mark; Van Gool, Luc; Williams, Christopher K. I.; Winn, John; Zisserman, Andrew
err分享
err收藏
Overcoming catastrophic forgetting in neural networks克服神经网络中的灾难性遗忘
err2017-03-14
err3.7K
errOAAI
errKirkpatricka, James; Pascanu, Razvan; Rabinowitz, Neil; Veness, Joel; Desjardins, Guillaume; Rusu, Andrei A.; Milan, Kieran; Quan, John; Ramalho, Tiago; Grabska-Barwinska, Agnieszka; Hassabis, Demis; Clopath, Claudia; Kumaran, Dharshan; Hadsell, Raia
err分享
err收藏
没有更多内容